# Fileset

[stgs-2s58.pdf](https://mdr.nims.go.jp/filesets/1376e0de-4a16-4fe2-a24c-eba287fd8e7d/download)

## Creator

Alperen Tüğen, Anna M. Seiler, Arthur Christianen, [Kenji Watanabe](https://orcid.org/0000-0003-3701-8119), [Takashi Taniguchi](https://orcid.org/0000-0002-1467-3105), Martin Kroner, Ataç İmamoğlu

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Optical Injection and Detection of Long-Lived Interlayer Excitons in van der Waals Heterostructures](https://mdr.nims.go.jp/datasets/0e2b54f4-36da-4042-ae27-3ed56979fc1f)

## Fulltext

Optical Injection and Detection of Long-Lived Interlayer Excitons in van der Waals HeterostructuresOptical Injection and Detection of Long-Lived Interlayer Excitons invan der Waals HeterostructuresAlperen Tüğen ,1,* Anna M. Seiler ,1,* Arthur Christianen,1 Kenji Watanabe ,2 Takashi Taniguchi ,3Martin Kroner,1 and Ataç İmamoğlu 11Institute for Quantum Electronics, ETH Zurich, CH-8093 Zurich, Switzerland2Research Center for Electronic and Optical Materials, National Institute for Materials Science, Tsukuba, Japan3Research Center for Materials Nanoarchitectonics, National Institute for Materials Science, Tsukuba, Japan(Received 6 June 2025; accepted 3 November 2025; published 10 December 2025)Interlayer excitons in semiconducting bilayers separated by insulating hexagonal boron nitride (h-BN)layers constitute a promising platform for investigation of strongly correlated bosonic phases. Here, wereport an optical method for the generation and characterization of long-lived interlayer excitons. Weconfirm the presence of tightly bound interlayer excitons by measuring 1s and 2s intralayer excitons in eachlayer concurrently. Using a pump-probe technique, we find interlayer exciton lifetimes up to 8.8 μs,increasing with the thickness of the h-BN. With optical access to long-lived interlayer excitons, ourapproach provides a new route to explore degenerate Bose–Fermi mixtures of excitons and itinerantelectrons with high spatial and temporal resolution.DOI: 10.1103/stgs-2s58Transition metal dichalcogenide (TMD) heterobilayershave recently emerged as a promising platform for creatinglong-lived interlayer excitons, bound electron-hole pairs,whose constituents reside in different layers. With theirlarge binding energy and a built-in out-of-plane dipole, theyoffer a route to investigating many-body phenomenaranging from superfluidity of dipolar excitons [1–4] toBose–Fermi mixtures exhibiting exotic electron pairingmechanisms [5,6] in van der Waals heterostructures.However, accessing these collective states depends oncontrolling the lifetime, density, and interactions of inter-layer excitons [3,7].Early demonstrations of long-lived interlayer excitonsreliedon theelectrical injectionof chargecarriers into separateTMDlayers [8–11].Capacitance spectroscopyandCoulomb-drag measurements provided compelling evidence for inter-layer exciton formation. However, these measurements offerno information on time dynamics and spatial distributions. Inparallel, optical pumping in aligned MoSe2–WSe2 bilayers,either in direct contact or separated by one or two hexagonalboron nitride (h-BN) layers [12–15], has been explored.Photoluminescence (PL) from such devices has revealedlifetimes up to 1.9 μs [16]. However, achieving longer life-times andhigher densities requires thickerh-BN spacer layersto suppress residual interlayer tunneling and radiative recom-bination [17]. This, in turn, significantly reduces PLemission,rendering direct optical measurements of the interlayerexciton dynamics untenable.Here, we demonstrate an optical scheme that overcomesthe limitations of both electrical and PL-based injection anddetection schemes. While inserting the h-BN spacer withup to seven layers to suppress interlayer tunneling, weinject high densities of electrons and holes into separateMoSe2 and WSe2 layers via nonresonant optical pumping,without requiring layer-selective contacts or complex gat-ing geometries [8,9]. Using time-resolved reflection spec-troscopy and the 2s exciton resonance as a spectroscopicprobe, we observe interlayer exciton formation and relax-ation dynamics in samples with negligible interlayerexciton PL. Under these conditions, we achieve interlayerexciton densities up to nIX ≈ 1012 cm−2 and lifetimes of8.8 μs. This combination of high density, long lifetime, andoptical control opens the door to studying strongly inter-acting interlayer exciton gases and exploring collectiveexcitonic phenomena.We investigate three TMD bilayer devices with varyingh-BN spacers [see Fig. 1(a)]. In the main text, we presentdata from Devices 1 and 2, both MoSe2–WSe2 bilayerswith h-BN spacers of 1- and 3–5-layers, while results fromDevice 3 (MoS2–WSe2, 5–7 layer h-BN spacer) areprovided in Supplemental Material (SM) [18]. In contrastto previous optical studies, all TMD layers within a deviceare angle-misaligned, suppressing interlayer hybridizationand resulting in negligible interlayer PL [28]. To determinethe charge configuration in the individual layers, weperform reflection spectroscopy. The upper panel ofFig. 1(b) shows spectra of the MoSe2 and WSe2 1s exciton*These authors contributed equally to this work.Published by the American Physical Society under the terms ofthe Creative Commons Attribution 4.0 International license.Further distribution of this work must maintain attribution tothe author(s) and the published article’s title, journal citation,and DOI.PHYSICAL REVIEW LETTERS 135, 246502 (2025)0031-9007=25=135(24)=246502(6) 246502-1 Published by the American Physical Societyhttps://orcid.org/0009-0008-8750-1006https://orcid.org/0000-0002-9883-9220https://orcid.org/0000-0003-3701-8119https://orcid.org/0000-0002-1467-3105https://orcid.org/0000-0002-0641-1631https://ror.org/05a28rw58https://ror.org/026v1ze26https://ror.org/026v1ze26https://crossmark.crossref.org/dialog/?doi=10.1103/stgs-2s58&domain=pdf&date_stamp=2025-12-10https://doi.org/10.1103/stgs-2s58https://doi.org/10.1103/stgs-2s58https://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/resonances as a function of the gate voltage Vμ (controls thechemical doping of the layers, see SM [18] for details),measured in Device 1. Upon injection of itinerant chargecarriers, the 1s exciton evolves into two distinct resonances[Fig. 1(b)]: the repulsive polaron (RP) and the attractivepolaron (AP) [29]. The left (right) side of the blue (red)dashed line in Fig. 1(b) corresponds to hole (electron)injection into the WSe2 (MoSe2) layer. The voltage rangebetween the dashed lines marks the charge-neutral regime,where neither layer is doped.Unlike previous experiments that bridged this energygap using large electric fields and an interlayer bias [8–11],we employ nonresonant optical pumping to inject chargecarriers into the TMD bilayer. We illuminate the samplewith a continuous-wave laser at 635 nm, above the bandgap [Fig. 1(c)] with pump and probe focused onto the samespot. Remarkably, as we show below, this optical dopingtechnique enables simultaneous electron- and hole-dopingin opposite layers even for devices with thicker h-BNspacers, overcoming the intrinsic band-offset limitationswithout the need for applying an electrical bias. Figure 1(d)displays the reflection spectra under continuous pumping atlaser power Ppump ¼ 6 μW. Here, the charge-neutral regionvanishes, and we observe simultaneous electron doping inthe MoSe2 layer and hole doping in the WSe2 layer. This isindicated by the reversal in the positions of the red and bluedashed lines [30].To visualize the overall charge configuration, Fig. 1(e)shows the sum of the peak 1s X (RP) reflection contrast ofthe MoSe2 and WSe2 layers normalized to the individualreflection contrast of each layer in the absence of charges(Zsum) as a function of Vμ and Ppump (see SM [18] fordetails). The yellow region labeled as (i, i) and outlined by ablack dotted contour denotes the approximate charge-neutral regime, where both excitons reach maximumcontrast. At higher Ppump, a dark blue region appears abovea light gray dotted contour line, indicating dual-layerdoping with opposite charge carriers (e, h). In the surround-ing regions, either WSe2 is hole doped (i, h) or MoSe2 iselectron doped (e, i), with the opposite layer remainingcharge neutral. Details on the asymmetry between hole andelectron doping are provided in SM [18].Next, we investigate whether the injected electrons andholes form bound interlayer exciton states or remainunbound. Prior theoretical work [31] demonstrated thatthe modification of the 1s exciton-polaron resonances canbe used to differentiate between bound and unboundelectron-hole pairs; however, the associated spectral shiftsAP(x2)RPXAP (x2)RP(x2)(x2)++probeprobe pump++ +(a)(c)(b)(d)(e)FIG. 1. (a) Schematic of the device structure. (b) Normalized reflection spectra (R=R0) of the MoSe2 and WSe2 layers as a function ofgate voltage Vμ. Resonances corresponding to the repulsive polaron (RP), attractive polaron (AP), and neutral exciton (X) are indicated.Blue and red dashed lines mark the onset of hole and electron doping, respectively. Aweak pump power was present during acquisition(Ppump ¼ 1.7 μW); its effect on the spectra is negligible. (c) Schematic under pump conditions. (d) R=R0 at finite pump power(Ppump ¼ 6 μW). (e) The sum of the peak 1s X (RP) reflection contrasts of the MoSe2 andWSe2 layers, each normalized to its respectivereflection contrast in the absence of charges, Zsum, is shown as a function of Vμ and Ppump. Distinct charge configurations are labeled as(i, i), (i, h), (e, i), and (h, e), representing combinations of intrinsic (i), hole (h), and electron (e) doping in the Mo and W layers,respectively. The dotted black contour marks the region where the normalized X (RP) reflection contrast in both layers exceeds 0.7,indicating approximate charge neutrality; note, however, that small but finite doping persists near the boundary of the (i, i) region. Thedotted light gray contour outlines the region with opposite doping in the two layers, where the normalized X (RP) reflection contrast ineach layer falls below 0.7. Arrows indicate the positions of vertical and horizontal linecuts. Data in panels (b)–(e) were acquired usingDevice 1.PHYSICAL REVIEW LETTERS 135, 246502 (2025)246502-2are rather small and can be masked by exciton linebroadening [9]. The origin of this relatively weak depend-ence is the small Bohr radius of 1s intralayer excitons,rendering them insensitive to the charges in the neighboringTMD layer [see the sketch in the upper panel of Fig. 2(a)].By contrast, Rydberg exciton states, such as the 2s state,exhibit a significantly larger Bohr radius [lower panel ofFig. 2(a)], making them more sensitive to the surroundingelectrostatic environment [32]. As such, the 2s excitonreflection contrast can be expected to respond differentlydepending on whether the charge carriers are (tightly)bound as interlayer excitons or remain free unboundparticles.Spectrally, we can resolve MoSe2 and WSe2 2s and 3sexcitons in the charge-neutral regime [Fig. 2(b)]. In the caseof free carriers residing in the same layer, 2s excitonsexhibit a blueshift and loss of oscillator strength even forsmall carrier densities. When charges are located in theopposite layer, the 2s-resonance still quickly disappears,but part of its oscillator strength is transferred to a newredshifted resonance which arises from the hybridizationwith the 2p-states [32–35].Interestingly, we observe that, in the regime where bothlayers are doped under finite Ppump, the MoSe2 and WSe22s exciton features persist with reduced intensity and only asmall spectral redshift (see SM [18]). To quantify thisbehavior across the charge map, we extract the WSe2 2scontrast, defined as the peak-to-peak amplitude of thedispersive Lorentzian lineshape associated with theWSe2 2s exciton [Fig. 2(c)]. Figures 2(d)–2(f) showdifferentiated spectral linecuts taken at different positionsin the phase diagram. In the (e, h) regime (black diamond),the 2s exciton contrast remains finite, even though thecharge densities match those in the red and blue diamondregions, where the 2s resonance vanishes. Moreover, the 2sresonance energy in the (e, h) region coincides with that atcharge neutrality (green diamond). This robustness of the2s resonance in the presence of dual-layer doping stands instark contrast to its drastic modification due to free chargesin either layer and points to a distinct phase: the formation(a) (b) (c)(d) (e) (f)FIG. 2. (a) Schematic of the 1s, 2s intralayer excitons and interlayer exciton (IX). (b) Differential reflection spectra (dðR=R0Þ=dE)plotted as a function of gate voltage Vμ, in the energy range of the Rydberg excitons. Data are shown for pump powers of Ppump ¼1.7 μW (upper panel) and Ppump ¼ 6 μW (lower panel). MoSe2 and WSe2 2s Rydberg excitons are labeled (see Fig. S7 [18] fordiscussion on 3s Rydberg excitons). (c) Color map of WSe2 2s contrast as a function of Vμ and Ppump. The dashed black and graycontour lines delineate different charge configurations, defined as in Fig. 1(e). The dashed pink contour marks the boundary where theWSe2 2s contrast falls below 0.055 (20%) in the region where the two layers are oppositely doped. (d),(e) Log-scale spectral linecuts atdistinct doping configurations [see diamond markers in panel (c)]. The blue and red curves correspond to only hole doping in WSe2 andonly electron doping in MoSe2, respectively, with the opposite layer remaining charge neutral. The black curve shows simultaneousdoping of both layers, with individual charge carrier densities matching those in the red and blue cases. (f) Differential reflection spectra(dðR=R0Þ=dE) plotted across the Rydberg exciton resonances. The green curve, measured at charge neutrality (green diamond), showsthe bare Rydberg excitons. The 2s resonance disappears when only one layer is doped (red and blue curves), but remains visible withouta significant spectral shift when both layers are doped simultaneously (black curve). Data in panels (b)–(f) were acquired using Device 1.PHYSICAL REVIEW LETTERS 135, 246502 (2025)246502-3of interlayer excitons. Because interlayer excitons arecharge neutral and do not generate strong in-plane electricfields, they do not directly hybridize the 2s and 2p-states.Instead, the leading-order interaction between the 2s andinterlayer excitons is a weak attractive van der Waalsinteraction (see SM [18]), leading to a small density-dependent redshift.We note that the persistence of the 2s was consistentlyobserved at different spots within a device as well as acrossmultiple devices, indicating that spatial inhomogeneitiesare unlikely to account for the observed feature. Since themeasured electron and hole trion binding energies in thetwo layers agree well with previously studied monolayerdevices, we can also rule out strong trapping of theoptically introduced charges [36] as an explanation forthe robustness of 2s excitons.We further observe that the 2s exciton contrast dimin-ishes progressively with increasing Ppump. This behavior isexpected: as the interlayer exciton spacing decreases andapproaches the 2s Bohr radius, the 2s state will cease to bebound. The fact that the 3s exciton resonance vanishes ateven lower interlayer exciton densities supports our explan-ation (Fig. S7 [18]). We emphasize that the 2s signalquenches at densities far below the expected interlayerexciton Mott density of ≈4 × 1012 cm−2 [8,31].Figures 3(a) and 3(b) show maps of the electron and holedensities, ne and nh, as functions of Vμ and Ppump,respectively (see SM [18]). From these data we extractthe minimum of ðne; nhÞ [minðne; nhÞ] and the excesscharge density ne − nh, plotted in Figs. 3(c) and 3(d),respectively. In the region where the 2s contrast is visibleand the excess charge approaches zero (pink dashedcontour), all carriers are paired and thus form interlayerexcitons. Thus, the interlayer exciton density is equal tominðne; nhÞ. Outside the pink contour, interlayer excitonsmay still be present alongside free carriers, but the 2sspectroscopic signature lacks sufficient contrast to confirmtheir presence. Here, minðne; nhÞ provides an upper boundon the attainable interlayer exciton density. In Device 1,where the MoSe2 and WSe2 layers are separated by amonolayer of h-BN, the minðne; nhÞ reaches up to3 × 1011 cm−2. By contrast, in Device 2, which featuresan h-BN spacer thickness of 3–5 layers, minðne; nhÞreaches 1 × 1012 cm−2 (Fig. S10 [18]).Our optical pumping scheme provides access to inter-layer exciton dynamics. We probe these dynamics usingtime-resolved reflection spectroscopy, with synchronizedmodulation of the pump and probe beams that allows forprogrammable delays between pulses (see SM [18] fordetails). In Fig. 4, we present the evolution of minðne; nhÞon a logarithmic scale as a function of pump–probe delay,enabling us to quantify the lifetime of the interlayer excitonstate at a representative point in the phase diagram, whereminðne; nhÞ is finite, the total excess charge is negligible,and the WSe2 2s exciton is present. The decay of theminðne; nhÞ signal with increasing time delay reflects therelaxation of interlayer excitons after the pump laser isturned off. Although the decay is clearly nonexponential,we extract a 1/e lifetime of 1.4 μs for Device 1, which(a) (b)(c) (d)FIG. 3. Charge carrier densities extracted from the opticalspectra measured in Device 1 as a function of the gate voltage,Vμ, and pump power, Ppump: (a) electron density, ne, in theMoSe2 layer; (b) hole density, nh, in theWSe2 layer; (c) minimumof ne and nh [minðne; nhÞ]; and (d) excess charge ne − nh. Dottedcontour lines denote the same boundaries depicted in Figs. 1(e)and 2(c).D1pump probetime delay0.5 μs5 μsD2FIG. 4. Time evolution of the minimum of the electron and holedensities, minðne; nhÞ, as a function of time delay between thepump and probe pulses (see inset). Data points from Device 1(D1, monolayer h-BN spacer) are shown as black filled circles,while those from Device 2 (D2, 3–5-layer h-BN spacer) areshown as dark red open circles. The dashed lines mark the densityat which minðne; nhÞ has decayed to 1=e of its initial value; thecorresponding delay times are 1.4 μs for Device 1 and 8.8 μs forDevice 2.PHYSICAL REVIEW LETTERS 135, 246502 (2025)246502-4contains a monolayer h-BN spacer. By contrast, a signifi-cantly longer lifetime of 8.8 μs is observed for Device 2,which has a thicker h-BN spacer of 3–5 layers.This prolonged lifetime is consistent with expectations:increased spacer thickness suppresses interlayer tunnelingand reduces the radiative decay rate, thereby extending theinterlayer exciton lifetime [16]. Surprisingly, the decaybecomes faster at lower carrier densities [37]. We speculatethat the shortened lifetime may stem from interlayerexcitons becoming more tightly bound as their density islowered: since the radiative decay rate scales inversely withthe square of the Bohr radius, we would expect the lifetimeto become shorter. Finally, we emphasize that our mea-surements cannot rule out a dominant, sample-dependent,nonradiative decay mechanism.The generation of interlayer excitons within a focusedoptical spot, combined with the ability to probe theirdynamics on picosecond timescales [38], provides a power-ful platform for exploring degenerate dipolar excitons withhigh spatial and temporal resolution. Crucially, it allows forspatially controlled exciton generation without the need forexternal electric fields, which are typically required inschemes requiring bias voltage for interlayer excitongeneration [8,9]. This capability opens the door to creatingmultiple excitonic reservoirs within a single device, pavingthe way for Josephson-like experiments [39–41]. In par-ticular, optically inducing high-density exciton populationsin two gate-defined traps [42,43] connected by a narrowchannel could provide a direct route to optically probing[44] coherent tunneling and phase dynamics betweencondensate regions.Acknowledgments—We thank Xiaobo Lu for fabricatingDevice 1. We thank Ivan Amelio, Haydn S. Adlong, andIgor Khanonkin for inspiring discussions. This work wassupported by the Swiss National Science Foundation(SNSF) under Grant No. 200021-204076. A. M. S. andA. C. acknowledge funding from an ETH PostdoctoralFellowship. K.W. and T. T. acknowledge support from theJSPS KAKENHI (Grants No. 21H05233 andNo. 23H02052), the CREST (JPMJCR24A5), JST, andWorld Premier International Research Center Initiative(WPI), MEXT, Japan.Data availability—The data that support the findings ofthis article are openly available [45].[1] M. Lu, N. Q. Burdick, S. H. Youn, and B. L. Lev, Phys. Rev.Lett. 107, 190401 (2011).[2] K. Aikawa, A. Frisch, M. Mark, S. Baier, A. Rietzler, R.Grimm, and F. Ferlaino, Phys. Rev. Lett. 108, 210401(2012).[3] M. Alloing, M. Beian, M. Lewenstein, D. Fuster, Y.González, L. González, R. Combescot, M. Combescot,and F. Dubin, Europhys. Lett. 107, 10012 (2014).[4] M. Combescot, R. Combescot, and F. Dubin, Rep. Prog.Phys. 80, 066501 (2017).[5] C. Zerba, C. Kuhlenkamp, A. Imamoǧlu, and M. Knap,Phys. Rev. Lett. 133, 056902 (2024).[6] J. von Milczewski, X. Chen, A. Imamoglu, and R. Schmidt,Phys. Rev. Lett. 133, 226903 (2024).[7] A. A. High, J. R. Leonard, A. T. Hammack, M. M. Fogler,L. V. Butov, A. V. Kavokin, K. L. Campman, and A. C.Gossard, Nature (London) 483, 584 (2012).[8] L. Ma, P. X. Nguyen, Z. Wang, Y. Zeng, K. Watanabe, T.Taniguchi, A. H. MacDonald, K. F. Mak, and J. Shan,Nature (London) 598, 585 (2021).[9] R. Qi, A. Y. Joe, Z. Zhang, Y. Zeng, T. Zheng, Q. Feng, J.Xie, E. Regan, Z. Lu, T. Taniguchi, K. Watanabe, S. Tongay,M. F. Crommie, A. H. MacDonald, and F. Wang, Nat.Commun. 14, 8264 (2023).[10] P. X. Nguyen, L. Ma, R. Chaturvedi, K. Watanabe, T.Taniguchi, J. Shan, and K. F. Mak, Science 388, 274(2025).[11] R. Qi, A. Y. Joe, Z. Zhang, J. Xie, Q. Feng, Z. Lu, Z. Wang,T. Taniguchi, K. Watanabe, S. Tongay, and F. Wang, Science388, 278 (2025).[12] P. Rivera, J. R. Schaibley, A. M. Jones, J. S. Ross, S. Wu, G.Aivazian, P. Klement, K. Seyler, G. Clark, N. J. Ghimireet al., Nat. Commun. 6, 6242 (2015).[13] L. A. Jauregui, A. Y. Joe, K. Pistunova, D. S. Wild, A. A.High, Y. Zhou, G. Scuri, K. De Greve, A. Sushko, C.-H. Yuet al., Science 366, 870 (2019).[14] Y. Shimazaki, I. Schwartz, K. Watanabe, T. Taniguchi, M.Kroner, and A. Imamoǧlu, Nature (London) 580, 472(2020).[15] F. Mahdikhanysarvejahany, D. N. Shanks, M. Klein, Q.Wang, M. R. Koehler, D. G. Mandrus, T. Taniguchi, K.Watanabe, O. L. Monti, B. J. LeRoy et al., Nat. Commun.13, 5354 (2022).[16] J. Cutshall, F. Mahdikhany, A. Roche, D. N. Shanks, M. R.Koehler, D. G. Mandrus, T. Taniguchi, K. Watanabe, Q.Zhu, B. J. LeRoy et al., Sci. Adv. 11, eadr1772 (2025).[17] D. Snoke, Adv. Condens. Matter Phys. 2011, 938609(2011).[18] See Supplemental Material at http://link.aps.org/supplemental/10.1103/stgs-2s58 for experimental details,additional figures and further discussion, which includesRefs. [19–27].[19] A. Reinhard, Strong photon-photon interactions in solidstate cavity QED, Ph.D. thesis, ETH Zurich, 2013.[20] J. G. Roch, D. Miserev, G. Froehlicher, N. Leisgang, L.Sponfeldner, K. Watanabe, T. Taniguchi, J. Klinovaja, D.Loss, and R. J. Warburton, Phys. Rev. Lett. 124, 187602(2020).[21] T. Wang, Z. Li, Z. Lu, Y. Li, S. Miao, Z. Lian, Y. Meng, M.Blei, T. Taniguchi, K. Watanabe et al., Phys. Rev. X 10,021024 (2020).[22] E. Liu, J. van Baren, T. Taniguchi, K. Watanabe, Y.-C.Chang, and C. H. Lui, Phys. Rev. B 99, 205420 (2019).[23] T. Smoleński, O. Cotlet, A. Popert, P. Back, Y. Shimazaki, P.Knüppel, N. Dietler, T. Taniguchi, K. Watanabe, M. Kroneret al., Phys. Rev. Lett. 123, 097403 (2019).[24] J. Kang, S. Tongay, J. Zhou, J. Li, and J. Wu, Appl. Phys.Lett. 102, 012111 (2013).PHYSICAL REVIEW LETTERS 135, 246502 (2025)246502-5https://doi.org/10.1103/PhysRevLett.107.190401https://doi.org/10.1103/PhysRevLett.107.190401https://doi.org/10.1103/PhysRevLett.108.210401https://doi.org/10.1103/PhysRevLett.108.210401https://doi.org/10.1209/0295-5075/107/10012https://doi.org/10.1088/1361-6633/aa50e3https://doi.org/10.1088/1361-6633/aa50e3https://doi.org/10.1103/PhysRevLett.133.056902https://doi.org/10.1103/PhysRevLett.133.226903https://doi.org/10.1038/nature10903https://doi.org/10.1038/s41586-021-03947-9https://doi.org/10.1038/s41467-023-43799-7https://doi.org/10.1038/s41467-023-43799-7https://doi.org/10.1126/science.adl1829https://doi.org/10.1126/science.adl1829https://doi.org/10.1126/science.adl1839https://doi.org/10.1126/science.adl1839https://doi.org/10.1038/ncomms7242https://doi.org/10.1126/science.aaw4194https://doi.org/10.1038/s41586-020-2191-2https://doi.org/10.1038/s41586-020-2191-2https://doi.org/10.1038/s41467-022-33082-6https://doi.org/10.1038/s41467-022-33082-6https://doi.org/10.1126/sciadv.adr1772https://doi.org/10.1155/2011/938609https://doi.org/10.1155/2011/938609http://link.aps.org/supplemental/10.1103/stgs-2s58http://link.aps.org/supplemental/10.1103/stgs-2s58http://link.aps.org/supplemental/10.1103/stgs-2s58http://link.aps.org/supplemental/10.1103/stgs-2s58http://link.aps.org/supplemental/10.1103/stgs-2s58https://doi.org/10.1103/PhysRevLett.124.187602https://doi.org/10.1103/PhysRevLett.124.187602https://doi.org/10.1103/PhysRevX.10.021024https://doi.org/10.1103/PhysRevX.10.021024https://doi.org/10.1103/PhysRevB.99.205420https://doi.org/10.1103/PhysRevLett.123.097403https://doi.org/10.1063/1.4774090https://doi.org/10.1063/1.4774090[25] P. Rivera, H. Yu, K. L. Seyler, N. P. Wilson, W. Yao, and X.Xu, Nat. Nanotechnol. 13, 1004 (2018).[26] Y. Yoon, Z. Zhang, R. Qi, A. Y. Joe, R. Sailus, K. Watanabe,T. Taniguchi, S. Tongay, and F. Wang, Nano Lett. 22, 10140(2022).[27] T. Handa, M. Holbrook, N. Olsen, L. N. Holtzman, L.Huber, H. I. Wang, M. Bonn, K. Barmak, J. C. Hone, A. N.Pasupathy et al., Sci. Adv. 10, eadj4060 (2024).[28] Device 1 shows very weak interlayer PL, whereas nointerlayer PL is observed in Device 2 or 3; see Fig. S2 inSupplemental Material [18].[29] M. Sidler, P. Back, O. Cotlet, A. Srivastava, T. Fink, M.Kroner, E. Demler, and A. Imamoglu, Nat. Phys. 13, 255(2017).[30] We emphasize that the two types of doping can be spectrallydistinguished in WSe2 [21]: the presence of a singleattractive polaron (AP) resonance in the WSe2 spectrumunequivocally indicates hole doping since electron dopingof WSe2 leads to two AP features associated with the singletand triplet trions. By contrast, MoSe2 displays a single APpeak under both electron and hole doping. Additionalconfirmation comes from Device 3, a MoS2=WSe2 hetero-structure, where the MoS2 layer clearly exhibits two APfeatures under electron doping, consistent with our inter-pretation of the spectra in Devices 1 and 2 (see Fig. S6 [18]).[31] I. Amelio, N. D. Drummond, E. Demler, R. Schmidt, and A.Imamoglu, Phys. Rev. B 107, 155303.[32] Y. Xu, S. Liu, D. A. Rhodes, K. Watanabe, T. Taniguchi, J.Hone, V. Elser, K. F. Mak, and J. Shan, Nature (London)587, 214 (2020).[33] A. Popert, Y. Shimazaki, M. Kroner, K. Watanabe, T.Taniguchi, A. Imamoǧlu, and T. Smoleński, Nano Lett.22, 7363 (2022).[34] J. Kim, H. Dery, and D. Van Tuan, Phys. Rev. B 112,L041301 (2025).[35] A. Christianen, A. M. Seiler, A. Tüǧen, and A. İmamoǧlu(to be published).[36] N. Kiper, H. S. Adlong, A. Christianen, M. Kroner, K.Watanabe, T. Taniguchi, and A. İmamoǧlu, Phys. Rev. X 15,011049 (2025).[37] The parabolic trend in the logarithmic plot in Fig. 4 isclearly visible; similar nonexponential time dependence hasalso been observed by Cutshall et al. [16] in a time-resolvedPL measurement.[38] T. Uto, B. Evrard, K. Watanabe, T. Taniguchi, M. Kroner,and A. İmamoǧlu, Phys. Rev. Lett. 132, 056901 (2024).[39] B. Khorana, Phys. Rev. 185, 299 (1969).[40] K. G. Lagoudakis, B. Pietka, M. Wouters, R. André, and B.Deveaud-Plédran, Phys. Rev. Lett. 105, 120403 (2010).[41] M. Abbarchi, A. Amo, V. Sala, D. Solnyshkov, H. Flayac, L.Ferrier, I. Sagnes, E. Galopin, A. Lemaître, G. Malpuechet al., Nat. Phys. 9, 275 (2013).[42] R. Rapaport, G. Chen, S. Simon, O. Mitrofanov, L. Pfeiffer,and P. M. Platzman, Phys. Rev. B 72, 075428 (2005).[43] A. Gärtner, L. Prechtel, D. Schuh, A. W. Holleitner, and J. P.Kotthaus, Phys. Rev. B 76, 085304 (2007).[44] E. C. Regan, D. Wang, C. Jin, M. I. Bakti Utama, B. Gao, X.Wei, S. Zhao, W. Zhao, Z. Zhang, K. Yumigeta et al., Nature(London) 579, 359 (2020).[45] A. Tüǧen andA.M. Seiler, 2025, 10.3929/ethz-c-000785878.PHYSICAL REVIEW LETTERS 135, 246502 (2025)246502-6https://doi.org/10.1038/s41565-018-0193-0https://doi.org/10.1021/acs.nanolett.2c04030https://doi.org/10.1021/acs.nanolett.2c04030https://doi.org/10.1126/sciadv.adj4060https://doi.org/10.1038/nphys3949https://doi.org/10.1038/nphys3949https://doi.org/10.1103/PhysRevB.107.155303https://doi.org/10.1038/s41586-020-2868-6https://doi.org/10.1038/s41586-020-2868-6https://doi.org/10.1021/acs.nanolett.2c02000https://doi.org/10.1021/acs.nanolett.2c02000https://doi.org/10.1103/xdcj-ckslhttps://doi.org/10.1103/xdcj-ckslhttps://doi.org/10.1103/PhysRevX.15.011049https://doi.org/10.1103/PhysRevX.15.011049https://doi.org/10.1103/PhysRevLett.132.056901https://doi.org/10.1103/PhysRev.185.299https://doi.org/10.1103/PhysRevLett.105.120403https://doi.org/10.1038/nphys2609https://doi.org/10.1103/PhysRevB.72.075428https://doi.org/10.1103/PhysRevB.76.085304https://doi.org/10.1038/s41586-020-2092-4https://doi.org/10.1038/s41586-020-2092-4https://doi.org/10.3929/ethz-c-000785878 Optical Injection and Detection of Long-Lived Interlayer Excitons in van der Waals Heterostructures Acknowledgments Data availability References